[OPE-L:5465] RE: Luxury goods and profit rate

Duncan K. Foley (dkf2@columbia.edu)
Mon, 15 Sep 1997 07:34:07 -0700 (PDT)

[ show plain text ]

In reply to Ajit's OPE-L:5393:

After discussing some other issues, Ajit replies to a suggestion of mine by

>Yes, I'm driving at the defense of w = pb, where b is given and p and w
>adjust to ensure b. This would be a good place to pick up the issue of
>'given' money wages in NS. You say NS would have no problem with w = pb
>where b is taken as given. Now, let us suppose we start off with a given b,
>which gives us a certain amount of w in terms of gold, which happens to be
>the money commodity. Now, assume that we change the money commodity from
>gold to silver, and adjust w in terms of silver such that workers can buy
>exactly b again. For the real economy nothing has changed. However, in the
>new solution approach this would in all likelihood lead to a change in the
>rate of exploitation, which would be an absurd idea.

I don't think this is true, since all prices and wages would change in the
proportion of the price of silver in terms of gold, so the ratio of the
wage bill to the value of the net product would not change, and the NI
would measure the same rate of surplus value.

This emphasizes the point that the monetary expression of labor time is
_not_ the labor embodied in the money commodity. I have a feeling that
there is some confusion on this point in the general debate about the NI.
(If prices were proportional to embodied labor coefficients, then the
monetary expression of labor time would be the inverse of the labor
embodied in the money commodity, but otherwise the two coefficients can
differ, because the implicit price of the money commodity in terms of other
commodities may be above or below the corresponding ratios of labor

>That's why I think the
>NS requires that one identifies a money commodity from the beginning and
>must fix the wages in terms of that money commodity. Otherwise, it cannot
>avoid the obverse side of the Ricardian problem. When Ricardo took any
>commodity as money commodity, he found that a change in real wages would
>change all the prices in such manners that it cannot be held that the size
>of the net output remains constant, which was absurd. And thus his search
>for the 'invariable measure of value'. So I think the problem of the
>'invariable measure of value' is very much a problem of the NS approach,
>which shows up in various different ways, and that Eatwell's contribution
>should be taken seriously by the NS people.

I'm afraid this has more to do with your projections onto the NS than
anything anybody actually wrote. The NS does not propose a theory of the
real or money wage, nor a theory of competitive pricing. As Dumenil and
Levy have tried to make clear, it is an interpretation of Marx's labor
theory of value that tries to make the theory operational in terms of the
actual statistics of real capitalist economies. The argument is that Marx
routinely translates labor time into money units, using a scalar
coefficient with the dimensions of units of money per unit of labor time,
which the NI calls the "monetary expression of labor time" . The NI argues
that, whatever other implications this may have, the monetary expression of
labor should minimally express the ratio of the money value of the net
product (value added) to the living labor expended to produce it. (I tried
to revive Marx's phrase "value of money", which he frequently uses in
exactly the sense of the ratio of labor time to money, but have run into a
lot of misunderstanding and criticism over the phrase, so I'm willing to go
with the more explicit MELT.) If you think, as I do, that the core of
Marx's labor theory of value is the idea that profit arises as the unpaid
labor of workers, and you want to maintain a strict quantitative connection
between profit and unpaid labor, then you are forced mathematically to
regard the wage divided by the MELT as the paid labor. The advantage is
that one can then answer questions like "what is the rate of surplus value
in the U.S. economy in 1997" directly, and without converting everything
into a parallel embodied labor coefficient accounting system which actually
depends on a whole lot of theoretical assumptions and the adequacy of
input-output tables.

Since this is only an interpretation of the theoretical categories (surplus
value, and so forth) it is compatible with any theory of the determination
of money or real wages, and any theory of pricing and profit rates
(including the Classical theory Marx focused on in Volume III of Capital,
which assumes that competition tends to equalize profit rates across

The motivation of the NS (which has many precursors, some of which I've
tracked down, and some of which I'm probably still unaware of) was to make
it possible to use Marx's categories in the analysis of real economic data.
I was particularly interested in the dynamics of the circuit of capital
theory as an alternative to production functions, for example. I don't see
how it gets one involved with an "invariable measure of value": in fact,
the MELT varies all the time in real economies.


Duncan K. Foley
Department of Economics
Barnard College
New York, NY 10027
fax: (212)-854-8947
e-mail: dkf2@columbia.edu